The 1000 Genomes Project, which ran between 2008 and 2015, is as close as it comes to a “catalogue of human variation.” The output of this initiative is a database of whole genome sequences from 26 distinct populations from around the world, all aligned to the same human reference sequence. This data is free to use, and is an excellent resource for researchers who want to study genetic variation in a gene across populations, but cannot afford to collect their own samples.
While it was active, The 1000 Genomes Project published their data in several phases; by the final phase (Phase 3), they had gathered samples from 2,504 individuals from the 26 targeted populations. In 2015, the International Genome Sample Resource (IGSR) was established to “ensure the future usability and accessibility of the 1000 Genomes data.” In keeping with this goal, the IGSR has: re-mapped the Phase 3 data to the latest two human reference sequences, GRCh37 and GRCh38, incorporated externally generated, published genomic data (such as RNA-seq data) into their own dataset, and begun adding previously unsampled populations the database.
Below is a map of the current populations represented in the 1000 Genomes dataset, as well as a reference list of the abbreviations used to identify these populations.
One last thing to note is that each of these populations falls under a “super population” which denotes the general area of the world each population is from. Many times, you will see information split up by these super populations instead of by each individual population. These super populations are as follows:
For the in-class labs this semester, each of you will be assigned a focal sub-population for our investigations into UCP1. Unless told otherwise, this is the population you’ll be tracking variation in for the remainder of our time investigating UCP1. You can find your assigned population here!
For more information about 1000 Genomes and IGSR, visit http://www.internationalgenome.org/home.
Bioinformatics is the “science of developing methods and software tools for collecting and understanding biological data.” It’s become huge academic and professional field in a relatively short time as big datasets proliferate in biology, thanks to rapid developments in sequencing technology and the advances in the various ‘-omics’ fields.
BU has an interdisciplinary Master’s Program in Bioinformatics, a Bioinformatics Research and Interdisciplinary Training Experience (BRITE REU) for undergraduate students, as well as a Bioinformatics Hub (the BUBHUB) meant to support faculty and students conducting research in bioinformatics. These might be good resources if you decide you like this kind of work and want to pursue it further.
Starting this semester, there’s even a student-led Biology/Bioinformatics Peer Coding Hour here at BU, where undergraduate and graduate students help each other with bioinformatics and statistical coding issues. The Peer Coding Hour will have their first meet-and-greet (with free food!) on October 2nd from 4-5pm in BRB 113. I recommend everyone join!
The 1000 Genomes Project, or even digitally recording the information DNA gives us, would not have been possible without this field. To understand the files that we will be working with (such as VCF files, which we will discuss later), it is beneficial to know how raw data is transformed in to digital information. In order to explain this process, I have included a simple flowchart that I will walk through.
The first step in this flowchart is the DNA sequencing itself. There are several kinds of sequencing, but we know from the 1000 Genomes paper we read that they used what is called an Illumina platform. Illumina uses a specific method of next-generation sequencing (NGS on the diagram). NGS is is a faster, more efficient, and more in-depth process of sequencing that is based on shearing the genome into small pieces and then reconstructing them en masse and in parallel (multiple times at once) using various proprietary technologies before mapping those pieces to a reference genome (typically the first, highest quality, or most completely sequenced individual genome of a species; this does not mean this is the representative, average, or most common version of the genome!). The proprietary Illumina platform was invented by Illumina, and uses a unique method of sequencing that makes it among the most efficient, affordable, and accurate ways of sequencing that we have today. BU has it’s own Illumina sequencing facility on campus. Illumina sequencing in itself is an incredibly complex process that we won’t talk about in detail here, but a good video that explains the process can be found here.
DNA sequence reads don’t come out of the machine nicely put together and cleaned, however. There are a few more steps to turn them in to nice, neat files. As shown in the diagram, the output of a sequencing machine is called a Fastq file. A Fastq file consists of a raw nucleotide sequence that is not yet aligned to a reference genome, and accompanying quality scores, which are scores that tell us how reliable the sequencing read for each base is. You can work with these files, but without aligning them to a reference genome we won’t be able to get as much from them as we want. That’s where the next step in the diagram comes in…
Alignment is the process of taking a chunk of DNA sequence and using a statistical algorithm to compare that chunk to a reference genome to figure out what portion of the genome that chunk represents. This is done with all the small sequence chunks that come from the initial Fastq file until you have a fully aligned genome. Once you have aligned your Fastq sequence to a reference sequence, you have a BAM file. A BAM file therefore not only contains an entire genome’s worth of genetic code, but also gives information about where any particular piece of code falls within the genome. These files are good to work with if you need an entire genome’s worth of information, or detailed information about every nucleotide in a region.
The final step in the flowchart is the VCF file, which is what we will be working with in our class. VCF files are the result of picking out just the variant nucleotide positions (in other words, loci where individual sequences differ from the reference) from a BAM file. Below, we will look at the VCF file type more in-depth, as we will be using VCF files in this class.
In our labs, we will be using VCF files to look at our candidate gene, UCP1. The VCF file format is a computer file format in which variant genetic information can be stored, as we have seen above. VCF files in particular are a way of formatting SNP-only information without having to deal with the rest of the sequence that the SNPs come from. Other file types, such as BAM files, have their own uses but for the purposes of our study (and most population genetics studies) they simply contain way more information than we need: a single BAM file containing an entire genome can be almost a terabyte (1000 gigabytes) in size!
VCF files are a text-file format which can be opened with a plain text editor on your computer, and can be analyzed using various softwares. Below I have included an example screenshot of what a VCF file looks like when opened in a plain text editor. This example compares what a simple representation of the sequence itself aligned to the reference (‘Alignment’) looks like in VCF format (‘VCF representation’):
As you can see from parts (b-g) of the figure, there is different notation that can be used depending on what type of SNP or variant position is being recorded. If you’re interested in more complex bioinformatic analyses with data like this, there’s more information about VCF files here.
Links to, and information for, all of the genome browsers that feature 1000 Genomes data is found here.
As the 1000 Genomes project was running, 1000 Genomes had its own “early access” genome browsers that allowed researchers to get detailed information about a specific gene of interest. These browsers, which are still available today, contain the open-access information that was updated with each phase of the project. However, these browsers are now outdated, and the most up-to-date genomic alignments for the 1000 Genomes project data are generated by Ensembl. Ensembl is a genome database that is maintained by the European Bioinformatics Instritute, and houses genomic data for many different species. Ensembl also has several different versions which are updated as new alignment information becomes available. For this class, we will be using the most up-to-date version of the Ensembl human geneome browser, the GRCh38 browser. Today we will use Ensembl to look at our gene of interest, UCP1.
Go to the website: http://useast.ensembl.org/Homo_sapiens/Info/Index
Find the search bar in the top left-hand corner and type in “UCP1.” Make sure the “category” drop-down menu is set to “Search all categories.” Click “go.”
Congratulations, you’ve found your gene! Now, let’s visually explore UCP1.
If you click on the “Go to region in detail” option directly above this image, you will get a more detailed version of the image. If you’re interested, you can do this on your own time, but we will not need to do that for the purposes of this class. What we will look at, however, is an image that shows all of the variants in UCP1.
Oh no! JUST IN THE LAST MONTH, Ensembl decided to retire this view of the human genome, as known variants have proliferated to the point where there’s TOO much information here. I’ll show you what it used to look like, just to give you an idea of how useful it was (note, you can still get this view for NON-human sequences).
In species where this still works, you will find an interactive image that looks like this:
You can click on each little variant box as well, which will give you information about each SNP, such as its rs ID number, its location, and what kind of mutation it is:
Feel free to play around with this image, but what will probably be more helpful is the variant table.
You will get a table that looks like this:
As you can see under the “Conseq. type” column, all of the variants at the top of the table are downstream variants.
Let’s do a little exercise, shall we? Above the table, there are some filtering options. Click on the “Consequences” filtering option and hit “Turn all off”
Now, turn back on all of the mutations that lie within the coding region of the gene AND will cause a change in genotype. Hint: I chose six types of mutation. After you’re done choosing, hit “apply changes.”
Now that we have a shortlist of SNPs that will actually come in handy for us, let’s learn how to get useful information about a specific SNP that we will use in later labs.
The first thing we’ll look is the Global Minor Allele Frequency (MAF). MAF is simply the frequency at which the second most common variant allele from the reference genome (i.e. the minor allele) occurs in a population. We will be able to get population breakdowns of the MAF, but first we’ll look at the Global MAF.
Under “Explore This Variant” there is several different tabs that will tell us different things about the SNP. In this class, we will be using two of these features in the future: “Population Genetics” and “Linkage Disequilibrium.” Right now, we will do a quick tour of these tabs so you can just see how they work.
First, click on the “Population Genetics” icon. You will get this page:
You can see in these pie charts that there are differences in the allele frequencies in each population.
Scroll down to the table that gives you the allele frequency breakdown for all the 1000 Genomes populations. Here is an example section of the 1000 Genomes data table:
As an exercise, find your population in this table, and make sure you can find the allele and genotype counts for your population. This will come in handy when we do our Hardy-Weinberg lab.
The final page we’ll explore is the “Linkage Disequilibrium” page.
Find your population. There are a few things we will look at.
First, click on the right-most “View Plot” link for your population. The icon is a reddish triangle. A few things will come up on this page.
The first thing you should see on the page is an image of Chromosome 4, with a red line marking where the SNP of interest is:
Think about how the location of this SNP on the chromosome will affect the likelihood of Linkage Disequilibrium. Is it more or less likely that this SNP will be out of Linkage Disequilibrium?
This is a Linkage Disequilibrium block. When we talk about Linkage Disequilibrium, we will learn how to read one of these blocks. For now, just know what it looks like.
Now, go back to the table on the previous page. For your population, click on the “Show” link under the “Variants in High LD” column. If your population has SNPs in high Linkage Disequilibrium, you will get a table that looks like this:
If your population DOESN’T have SNPs in high Linkage Disequilibrium, think about what that tells us about the importance of UCP1 in your population.
Now that we have explored some important features of the Ensembl website, we can learn how to download some data of our own!
Usually when you are working with genomic information, you are given a whole chromosome or even a whole genome’s worth of information in either a BAM file or a VCF file. If you only need to look at one part of the genome, it can be very annoying to work with a lot of extra data. The Ensembl data slicer is a convenient way to get only the amount of data that you want without using a programto cut it out yourself. Therefore, we will use this tool to get the data for our analysis of UCP1. We will be taking two slices of data today, one that contains all of the SNPs in UCP1, and one that contains only about 1/4 of the SNPs in the gene. We will be using the larger slice in most of our analyses, but we will need the smaller slice for Lab 3.
One thing to note about the data slicer is that it is only available for the GRCh37 version of Ensembl, which is one version before GRCh38 (the version we have been using). What’s different about the two versions of this site is that all of the 1000 Genomes data is aligned to a different reference genome in each version. All this means for us is that UCP1 will have different genomic coordinates in each version. Similarly, the coordinates of SNPs will be affected, but the SNP ID numbers will be the same since those are universal. Don’t worry about finding the coordinates of UCP1, I will give the GRCh37 to you here.
The link to the data slicer is available here: http://grch37.ensembl.org/Homo_sapiens/Tools/DataSlicer?db=core.
Now, on to the tutorial!
First off, the file format should be set for VCF. If it’s not, click the drop-down menu and select VCF.
In the “region lookup” bar, copy and paste in the location 4:141,480,588-141,489,959. These are the GRCh37 version alignment coordinates for the gene UCP1. This is the larger of the two chunks we will be taking.
In in the “Choose data file” dropdown list, make sure “phase 3” is selected. This will ensure you get data from the last phase of the 1000 Genomes project.
In the “filters” category, select “By populations.” This will give you a dropdown menu of all of the 1000 Genomes populations. Select the population that you were assigned, so that you only get the data for that population.
The filled-in interface should look like this:
Hit the “run” button at the bottom of the page.
At the top of the page, hit “New Job” and repeat this process, but this time with the coordinates 4:141481462-141485260. These are the coordinates for the smaller slice that we will be taking.
When you have clicked “run” for both jobs, you will see this table will pop up, and will tell you when your job has been processed. Click “View results” to look at your results.
We check our file to see if the body is there because sometimes the server will malfunction and give you only the head of the VCF file. If that happens, repeat the data slicer process. Check both files in this same way.
The last thing we’ll do is save our file and put our newly downloaded files in to the folders we made in the SCC tutorial with Charlie Jahnke. To do that:
After you drag and drop your files, you’ll see them in your directory:
And now, you’re good to go for Lab 2!